COX REGRESSION FOR AUERRODS and LOG(WBC) -- YOURNAME 1 SURVIVAL TIMES for initial white blood-cell count and Auer cell type THE DATA AS SAS SEES IT 21:58 Monday, November 14, 2005 Note the different dependence on Logwbc in the two groups Log Obs rodtype havenorods weeks Logwbc Intwbc 1 NoRods 1 2 4.43136 4.43136 2 NoRods 1 3 4.00000 4.00000 3 NoRods 1 3 4.32222 4.32222 4 NoRods 1 3 4.44716 4.44716 5 NoRods 1 4 4.27875 4.27875 6 NoRods 1 4 4.41497 4.41497 7 NoRods 1 4 5.00000 5.00000 8 NoRods 1 7 3.17609 3.17609 9 NoRods 1 8 4.49136 4.49136 10 NoRods 1 16 3.95424 3.95424 11 NoRods 1 17 3.60206 3.60206 12 NoRods 1 22 3.72428 3.72428 13 NoRods 1 30 4.89763 4.89763 14 NoRods 1 43 5.00000 5.00000 15 NoRods 1 56 3.64345 3.64345 16 NoRods 1 65 3.47712 3.47712 17 WithRods 0 1 5.00000 -5.00000 18 WithRods 0 1 5.00000 -5.00000 19 WithRods 0 4 4.23045 -4.23045 20 WithRods 0 5 3.71600 -3.71600 21 WithRods 0 16 3.77815 -3.77815 22 WithRods 0 22 4.54407 -4.54407 23 WithRods 0 26 4.50515 -4.50515 24 WithRods 0 39 3.73239 -3.73239 25 WithRods 0 56 3.97313 -3.97313 26 WithRods 0 65 3.36173 -3.36173 27 WithRods 0 65 5.00000 -5.00000 28 WithRods 0 100 3.63347 -3.63347 29 WithRods 0 108 4.02119 -4.02119 30 WithRods 0 121 4.00000 -4.00000 31 WithRods 0 134 3.41497 -3.41497 32 WithRods 0 143 3.84510 -3.84510 33 WithRods 0 156 2.87506 -2.87506 COX REGRESSION FOR AUERRODS and LOG(WBC) -- YOURNAME 2 SURVIVAL TIMES for initial white blood-cell count and Auer cell type PHREG FOR WEEKS ON RODTYPE AND LOGWBC 21:58 Monday, November 14, 2005 The PHREG Procedure Model Information Data Set WORK.PHWBC Dependent Variable weeks Ties Handling BRESLOW Number of Observations Read 33 Number of Observations Used 33 Summary of the Number of Event and Censored Values Percent Total Event Censored Censored 33 33 0 0.00 Convergence Status Convergence criterion (GCONV=1E-8) satisfied. Model Fit Statistics Without With Criterion Covariates Covariates -2 LOG L 171.994 158.601 AIC 171.994 162.601 SBC 171.994 165.594 Testing Global Null Hypothesis: BETA=0 Test Chi-Square DF Pr > ChiSq Likelihood Ratio 13.3925 2 0.0012 Score 14.0650 2 0.0009 Wald 12.9674 2 0.0015 Analysis of Maximum Likelihood Estimates Parameter Standard Hazard Variable DF Estimate Error Chi-Square Pr > ChiSq Ratio havenorods 1 1.00552 0.42442 5.6130 0.0178 2.733 Logwbc 1 0.75453 0.31090 5.8900 0.0152 2.127 COX REGRESSION FOR AUERRODS and LOG(WBC) -- YOURNAME 3 SURVIVAL TIMES for initial white blood-cell count and Auer cell type PHREG FOR WEEKS ON RODTYPE, LOGWBC, AND INTERACTION 21:58 Monday, November 14, 2005 The PHREG Procedure Model Information Data Set WORK.PHWBC Dependent Variable weeks Ties Handling BRESLOW Number of Observations Read 33 Number of Observations Used 33 Summary of the Number of Event and Censored Values Percent Total Event Censored Censored 33 33 0 0.00 Convergence Status Convergence criterion (GCONV=1E-8) satisfied. Model Fit Statistics Without With Criterion Covariates Covariates -2 LOG L 171.994 156.184 AIC 171.994 162.184 SBC 171.994 166.674 Testing Global Null Hypothesis: BETA=0 Test Chi-Square DF Pr > ChiSq Likelihood Ratio 15.8098 3 0.0012 Score 14.4508 3 0.0024 Wald 12.4656 3 0.0059 Analysis of Maximum Likelihood Estimates Parameter Standard Hazard Variable DF Estimate Error Chi-Square Pr > ChiSq Ratio havenorods 1 5.07089 2.67963 3.5811 0.0584 159.315 Logwbc 1 0.81644 0.31392 6.7642 0.0093 2.262 LogIntwbc 1 -0.48434 0.31383 2.3818 0.1228 0.616 COX REGRESSION FOR AUERRODS and LOG(WBC) -- YOURNAME 4 SURVIVAL TIMES for initial white blood-cell count and Auer cell type SEPARATE SLOPES FOR EACH TREATMENT: THE DATA AS SAS SEES IT 21:58 Monday, November 14, 2005 Obs rodtype havenorods weeks Logwbc logwbcwithrods logwbcnorods 1 NoRods 1 2 4.43136 0.00000 4.43136 2 NoRods 1 3 4.00000 0.00000 4.00000 3 NoRods 1 3 4.32222 0.00000 4.32222 4 NoRods 1 3 4.44716 0.00000 4.44716 5 NoRods 1 4 4.27875 0.00000 4.27875 6 NoRods 1 4 4.41497 0.00000 4.41497 7 NoRods 1 4 5.00000 0.00000 5.00000 8 NoRods 1 7 3.17609 0.00000 3.17609 9 NoRods 1 8 4.49136 0.00000 4.49136 10 NoRods 1 16 3.95424 0.00000 3.95424 11 NoRods 1 17 3.60206 0.00000 3.60206 12 NoRods 1 22 3.72428 0.00000 3.72428 13 NoRods 1 30 4.89763 0.00000 4.89763 14 NoRods 1 43 5.00000 0.00000 5.00000 15 NoRods 1 56 3.64345 0.00000 3.64345 16 NoRods 1 65 3.47712 0.00000 3.47712 17 WithRods 0 1 5.00000 5.00000 0.00000 18 WithRods 0 1 5.00000 5.00000 0.00000 19 WithRods 0 4 4.23045 4.23045 0.00000 20 WithRods 0 5 3.71600 3.71600 0.00000 21 WithRods 0 16 3.77815 3.77815 0.00000 22 WithRods 0 22 4.54407 4.54407 0.00000 23 WithRods 0 26 4.50515 4.50515 0.00000 24 WithRods 0 39 3.73239 3.73239 0.00000 25 WithRods 0 56 3.97313 3.97313 0.00000 26 WithRods 0 65 3.36173 3.36173 0.00000 27 WithRods 0 65 5.00000 5.00000 0.00000 28 WithRods 0 100 3.63347 3.63347 0.00000 29 WithRods 0 108 4.02119 4.02119 0.00000 30 WithRods 0 121 4.00000 4.00000 0.00000 31 WithRods 0 134 3.41497 3.41497 0.00000 32 WithRods 0 143 3.84510 3.84510 0.00000 33 WithRods 0 156 2.87506 2.87506 0.00000 COX REGRESSION FOR AUERRODS and LOG(WBC) -- YOURNAME 5 SURVIVAL TIMES for initial white blood-cell count and Auer cell type COX PH REGRESSION with SEPARATE SLOPES for each rod type 21:58 Monday, November 14, 2005 The PHREG Procedure Model Information Data Set WORK.PHWBC Dependent Variable weeks Ties Handling BRESLOW Number of Observations Read 33 Number of Observations Used 33 Summary of the Number of Event and Censored Values Percent Total Event Censored Censored 33 33 0 0.00 Convergence Status Convergence criterion (GCONV=1E-8) satisfied. Model Fit Statistics Without With Criterion Covariates Covariates -2 LOG L 171.994 156.184 AIC 171.994 162.184 SBC 171.994 166.674 Testing Global Null Hypothesis: BETA=0 Test Chi-Square DF Pr > ChiSq Likelihood Ratio 15.8098 3 0.0012 Score 14.4508 3 0.0024 Wald 12.4656 3 0.0059 Analysis of Maximum Likelihood Estimates Parameter Standard Hazard Variable DF Estimate Error Chi-Square Pr > ChiSq Ratio havenorods 1 5.07089 2.67963 3.5811 0.0584 159.315 logwbcwithrods 1 1.30078 0.47385 7.5357 0.0060 3.672 logwbcnorods 1 0.33210 0.41174 0.6505 0.4199 1.394 COX REGRESSION FOR AUERRODS and LOG(WBC) -- YOURNAME 6 SURVIVAL TIMES for initial white blood-cell count and Auer cell type STRATIFIED COX PH REGRESSION by rod type 21:58 Monday, November 14, 2005 rodtype=NoRods The PHREG Procedure Model Information Data Set WORK.PHWBC Dependent Variable weeks Ties Handling BRESLOW Number of Observations Read 16 Number of Observations Used 16 Summary of the Number of Event and Censored Values Percent Total Event Censored Censored 16 16 0 0.00 Convergence Status Convergence criterion (GCONV=1E-8) satisfied. Model Fit Statistics Without With Criterion Covariates Covariates -2 LOG L 62.307 61.456 AIC 62.307 63.456 SBC 62.307 64.229 Testing Global Null Hypothesis: BETA=0 Test Chi-Square DF Pr > ChiSq Likelihood Ratio 0.8502 1 0.3565 Score 0.8638 1 0.3527 Wald 0.8500 1 0.3565 Analysis of Maximum Likelihood Estimates Parameter Standard Hazard Variable DF Estimate Error Chi-Square Pr > ChiSq Ratio Logwbc 1 0.40222 0.43626 0.8500 0.3565 1.495 COX REGRESSION FOR AUERRODS and LOG(WBC) -- YOURNAME 7 SURVIVAL TIMES for initial white blood-cell count and Auer cell type STRATIFIED COX PH REGRESSION by rod type 21:58 Monday, November 14, 2005 rodtype=WithRods The PHREG Procedure Model Information Data Set WORK.PHWBC Dependent Variable weeks Ties Handling BRESLOW Number of Observations Read 17 Number of Observations Used 17 Summary of the Number of Event and Censored Values Percent Total Event Censored Censored 17 17 0 0.00 Convergence Status Convergence criterion (GCONV=1E-8) satisfied. Model Fit Statistics Without With Criterion Covariates Covariates -2 LOG L 67.398 60.457 AIC 67.398 62.457 SBC 67.398 63.290 Testing Global Null Hypothesis: BETA=0 Test Chi-Square DF Pr > ChiSq Likelihood Ratio 6.9416 1 0.0084 Score 7.1982 1 0.0073 Wald 6.7961 1 0.0091 Analysis of Maximum Likelihood Estimates Parameter Standard Hazard Variable DF Estimate Error Chi-Square Pr > ChiSq Ratio Logwbc 1 1.23730 0.47462 6.7961 0.0091 3.446 COX REGRESSION FOR AUERRODS and LOG(WBC) -- YOURNAME 8 SURVIVAL TIMES for initial white blood-cell count and Auer cell type CHECKING TIME-DEPENDENCE OF RELATIVE LOGWBC HAZARD USING THE CORRECT TIME (FACTORTIME) 21:58 Monday, November 14, 2005 STRATIFIED COX REGRESSION FOR EACH ROD TYPE rodtype=NoRods The PHREG Procedure Model Information Data Set WORK.PHWBC Dependent Variable weeks Ties Handling BRESLOW Number of Observations Read 16 Number of Observations Used 16 Summary of the Number of Event and Censored Values Percent Total Event Censored Censored 16 16 0 0.00 Convergence Status Convergence criterion (GCONV=1E-8) satisfied. Model Fit Statistics Without With Criterion Covariates Covariates -2 LOG L 62.307 60.920 AIC 62.307 64.920 SBC 62.307 66.465 Testing Global Null Hypothesis: BETA=0 Test Chi-Square DF Pr > ChiSq Likelihood Ratio 1.3867 2 0.4999 Score 1.4297 2 0.4893 Wald 1.2561 2 0.5336 COX REGRESSION FOR AUERRODS and LOG(WBC) -- YOURNAME 9 SURVIVAL TIMES for initial white blood-cell count and Auer cell type CHECKING TIME-DEPENDENCE OF RELATIVE LOGWBC HAZARD USING THE CORRECT TIME (FACTORTIME) 21:58 Monday, November 14, 2005 STRATIFIED COX REGRESSION FOR EACH ROD TYPE rodtype=NoRods The PHREG Procedure Analysis of Maximum Likelihood Estimates Parameter Standard Hazard Variable DF Estimate Error Chi-Square Pr > ChiSq Ratio Logwbc 1 0.05763 0.64263 0.0080 0.9285 1.059 factortime 1 0.02549 0.03608 0.4988 0.4800 1.026 COX REGRESSION FOR AUERRODS and LOG(WBC) -- YOURNAME 10 SURVIVAL TIMES for initial white blood-cell count and Auer cell type CHECKING TIME-DEPENDENCE OF RELATIVE LOGWBC HAZARD USING THE CORRECT TIME (FACTORTIME) 21:58 Monday, November 14, 2005 STRATIFIED COX REGRESSION FOR EACH ROD TYPE rodtype=WithRods The PHREG Procedure Model Information Data Set WORK.PHWBC Dependent Variable weeks Ties Handling BRESLOW Number of Observations Read 17 Number of Observations Used 17 Summary of the Number of Event and Censored Values Percent Total Event Censored Censored 17 17 0 0.00 Convergence Status Convergence criterion (GCONV=1E-8) satisfied. Model Fit Statistics Without With Criterion Covariates Covariates -2 LOG L 67.398 60.445 AIC 67.398 64.445 SBC 67.398 66.112 Testing Global Null Hypothesis: BETA=0 Test Chi-Square DF Pr > ChiSq Likelihood Ratio 6.9533 2 0.0309 Score 7.1999 2 0.0273 Wald 6.7766 2 0.0338 COX REGRESSION FOR AUERRODS and LOG(WBC) -- YOURNAME 11 SURVIVAL TIMES for initial white blood-cell count and Auer cell type CHECKING TIME-DEPENDENCE OF RELATIVE LOGWBC HAZARD USING THE CORRECT TIME (FACTORTIME) 21:58 Monday, November 14, 2005 STRATIFIED COX REGRESSION FOR EACH ROD TYPE rodtype=WithRods The PHREG Procedure Analysis of Maximum Likelihood Estimates Parameter Standard Hazard Variable DF Estimate Error Chi-Square Pr > ChiSq Ratio Logwbc 1 1.18117 0.70488 2.8079 0.0938 3.258 factortime 1 0.00129 0.01203 0.0115 0.9144 1.001 COX REGRESSION FOR AUERRODS and LOG(WBC) -- YOURNAME 12 SURVIVAL TIMES for initial white blood-cell count and Auer cell type CHECKING TIME-DEPENDENCE OF RELATIVE LOGWBC HAZARD USING THE INCORRECT `RECORDIME' INSTEAD OF `FACTORTIME': THIS OUTPUT IS INCORRECT: NOTE THE DIFFERENCE IN OUTPUT! 21:58 Monday, November 14, 2005 rodtype=NoRods The PHREG Procedure Model Information Data Set WORK.PHBAD Dependent Variable weeks Ties Handling BRESLOW Number of Observations Read 16 Number of Observations Used 16 Summary of the Number of Event and Censored Values Percent Total Event Censored Censored 16 16 0 0.00 Convergence Status Convergence criterion (GCONV=1E-8) satisfied. Model Fit Statistics Without With Criterion Covariates Covariates -2 LOG L 62.307 19.511 AIC 62.307 23.511 SBC 62.307 25.056 Testing Global Null Hypothesis: BETA=0 Test Chi-Square DF Pr > ChiSq Likelihood Ratio 42.7956 2 <.0001 Score 15.7937 2 0.0004 Wald 4.2489 2 0.1195 COX REGRESSION FOR AUERRODS and LOG(WBC) -- YOURNAME 13 SURVIVAL TIMES for initial white blood-cell count and Auer cell type CHECKING TIME-DEPENDENCE OF RELATIVE LOGWBC HAZARD USING THE INCORRECT `RECORDIME' INSTEAD OF `FACTORTIME': THIS OUTPUT IS INCORRECT: NOTE THE DIFFERENCE IN OUTPUT! 21:58 Monday, November 14, 2005 rodtype=NoRods The PHREG Procedure Analysis of Maximum Likelihood Estimates Parameter Standard Hazard Variable DF Estimate Error Chi-Square Pr > ChiSq Ratio Logwbc 1 3.70414 2.01207 3.3891 0.0656 40.615 recordtime 1 -0.51621 0.25051 4.2463 0.0393 0.597 COX REGRESSION FOR AUERRODS and LOG(WBC) -- YOURNAME 14 SURVIVAL TIMES for initial white blood-cell count and Auer cell type CHECKING TIME-DEPENDENCE OF RELATIVE LOGWBC HAZARD USING THE INCORRECT `RECORDIME' INSTEAD OF `FACTORTIME': THIS OUTPUT IS INCORRECT: NOTE THE DIFFERENCE IN OUTPUT! 21:58 Monday, November 14, 2005 rodtype=WithRods The PHREG Procedure Model Information Data Set WORK.PHBAD Dependent Variable weeks Ties Handling BRESLOW Number of Observations Read 17 Number of Observations Used 17 Summary of the Number of Event and Censored Values Percent Total Event Censored Censored 17 17 0 0.00 Convergence Status Convergence criterion (GCONV=1E-8) satisfied. Model Fit Statistics Without With Criterion Covariates Covariates -2 LOG L 67.398 16.207 AIC 67.398 20.207 SBC 67.398 21.874 Testing Global Null Hypothesis: BETA=0 Test Chi-Square DF Pr > ChiSq Likelihood Ratio 51.1910 2 <.0001 Score 26.5307 2 <.0001 Wald 6.8354 2 0.0328 COX REGRESSION FOR AUERRODS and LOG(WBC) -- YOURNAME 15 SURVIVAL TIMES for initial white blood-cell count and Auer cell type CHECKING TIME-DEPENDENCE OF RELATIVE LOGWBC HAZARD USING THE INCORRECT `RECORDIME' INSTEAD OF `FACTORTIME': THIS OUTPUT IS INCORRECT: NOTE THE DIFFERENCE IN OUTPUT! 21:58 Monday, November 14, 2005 rodtype=WithRods The PHREG Procedure Analysis of Maximum Likelihood Estimates Parameter Standard Hazard Variable DF Estimate Error Chi-Square Pr > ChiSq Ratio Logwbc 1 4.65721 1.86347 6.2461 0.0124 105.342 recordtime 1 -0.06713 0.02582 6.7599 0.0093 0.935 COX REGRESSION FOR AUERRODS and LOG(WBC) -- YOURNAME 16 SURVIVAL TIMES for initial white blood-cell count and Auer cell type CORRECT MODELING FOR CHECKING NO-ROD TIME-DEPENDENCE 21:58 Monday, November 14, 2005 The PHREG Procedure Model Information Data Set WORK.PHWBC Dependent Variable weeks Ties Handling BRESLOW Number of Observations Read 33 Number of Observations Used 33 Summary of the Number of Event and Censored Values Percent Total Event Censored Censored 33 33 0 0.00 Convergence Status Convergence criterion (GCONV=1E-8) satisfied. Model Fit Statistics Without With Criterion Covariates Covariates -2 LOG L 171.994 155.822 AIC 171.994 163.822 SBC 171.994 169.808 Testing Global Null Hypothesis: BETA=0 Test Chi-Square DF Pr > ChiSq Likelihood Ratio 16.1716 4 0.0028 Score 15.7615 4 0.0034 Wald 12.9943 4 0.0113 Analysis of Maximum Likelihood Estimates Parameter Standard Hazard Variable DF Estimate Error Chi-Square Pr > ChiSq Ratio havenorods 1 4.54699 2.84322 2.5576 0.1098 94.348 logwbcwithrods 1 1.29793 0.47794 7.3748 0.0066 3.662 logwbcnorods 1 0.39372 0.42771 0.8474 0.3573 1.482 factortime 1 0.01152 0.01889 0.3720 0.5419 1.012